Real-time Sign Language Recognition using Machine Learning and Neural Network

Languages based on gestures have recently garnered a lot of interest. In sign language, they are one of the nonverbal communication modalities. It is most commonly used by deaf and dumb individuals with vocal or hearing issues in order to convey and talk with one other and normal people without disabilities. The motion, form, and location of hands, along with the body gestures and facial expressions, are used for conveying the meaning in American Sign Language(ASL). ASL translator also aids in the communication between signers and non-signers. The most efficient way in which they can convey and communicate is through using gestures and that is particularly known as Sign language. Hearing-impaired persons find it very complex and a problem to communicate with the normal people that are in the rest of the world, and as a result, they are isolated, that is the humans with speech impediments communicate via gestures, which makes it difficult for them to mix with the able-bodied. Hence, an interpretive system is required to serve as a link between them and individuals who do not understand their sign language. To assist them overcome these obstacles, we provide a project that will be beneficial to them. This project proposes a Machine Learning-based method for recognising hand gestures and translating them into speech/text, as well as the reverse, converting speech/text to gesture.It introduces a conceptual framework that can interpret sign language autonomously, allowing deaf and dumb individuals to convey more readily with one another and with regular people. Recognition systems and motion tracking are two important and crucial areas of study and are still in their infancy.